Fact-checked by Grok 2 weeks ago

Fuzzy mathematics

Fuzzy mathematics is a branch of mathematics that extends classical mathematical frameworks to handle imprecision, uncertainty, and vagueness by allowing elements to have degrees of membership between 0 and 1, rather than binary inclusion. Introduced by Lotfi A. Zadeh in his 1965 paper "Fuzzy Sets," it generalizes set theory through the concept of fuzzy sets, where membership functions assign partial belonging to elements in a universe. This foundational idea enables the representation of gradual transitions and subjective judgments, contrasting with crisp sets that enforce strict boundaries.^[1] The field encompasses fuzzy set theory as its core, along with fuzzy logic, which adapts logical operations to multivalued truth degrees for approximate reasoning.^[2] Key operations include fuzzy intersection (often using the minimum), union (maximum), and complement (1 minus membership), which can be parameterized for flexibility in applications.^[1] Fuzzy mathematics has branched into numerous subareas, such as fuzzy algebra for generalized group and ring structures, fuzzy topology for open sets with graded memberships, fuzzy calculus for derivatives under uncertainty, fuzzy graphs for relational modeling with ambiguity, and fuzzy differential equations for dynamic systems with imprecise parameters.^[3] Since its inception, fuzzy mathematics has evolved to integrate with other disciplines, including hybrid systems like adaptive neuro-fuzzy inference systems (ANFIS) that combine fuzzy logic with neural networks for enhanced learning and control.^[2] It finds extensive applications in engineering fields such as control systems (e.g., fuzzy controllers for stabilization), artificial intelligence for decision-making under uncertainty, robotics for navigation in vague environments, and power systems for fault diagnostics.^[1] These developments underscore its role in bridging mathematical rigor with real-world imprecision, influencing areas from biomedical engineering to data analysis.^[2]

Foundations

Definition

Fuzzy mathematics is a branch of mathematics designed to handle uncertainty, imprecision, and vagueness by extending classical mathematical structures to incorporate degrees of truth and membership, rather than relying solely on binary distinctions such as true/false or included/excluded. This framework models real-world phenomena where boundaries are not sharply defined, such as concepts like "tall" or "hot," allowing for a continuum of values to represent partial belonging or partial truth.^[4] The foundational idea was introduced by Lotfi A. Zadeh in 1965 with the development of fuzzy set theory, which generalizes classical set theory by permitting elements to have membership grades ranging continuously from 0 to 1. In this approach, a fuzzy set A in a universe X is defined as A = \{(x, \mu_A(x)) \mid x \in X\}, where \mu_A(x) \in [0,1] denotes the degree of membership of x in A. This contrasts sharply with crisp sets in classical mathematics, where membership is strictly binary: \mu(x) = 1 if x belongs to the set and \mu(x) = 0 otherwise.^[5] The scope of fuzzy mathematics encompasses fuzzy set theory as its core, along with fuzzy logic and related structures, enabling applications across mathematics, engineering, and computational intelligence to address problems involving incomplete or ambiguous information.^[4]

Historical Development

Fuzzy mathematics originated with Lotfi A. Zadeh's seminal 1965 paper "Fuzzy Sets," published in the journal Information and Control, which introduced the concept to address linguistic vagueness and limitations of classical binary logic in handling imprecise information. Zadeh was motivated by challenges in pattern recognition and natural language processing, where traditional set theory failed to capture gradations of membership.^[6] The roots of fuzzy sets trace back to earlier developments in multivalued logics, notably Jan Łukasiewicz's work in the 1920s on three-valued and infinite-valued logics to manage indeterminacy in propositions, as well as connections to probability theory for uncertainty modeling, though Zadeh provided the formal framework for fuzzy sets as a distinct mathematical structure.^[7] In the 1970s, Zadeh expanded the theory into fuzzy logic, with key publications like his 1975 paper on fuzzy logic and approximate reasoning, enabling computational handling of vagueness. This period also saw practical advancements, such as Ebrahim Mamdani's 1975 development of fuzzy logic controllers for industrial applications, demonstrated through linguistic synthesis for a steam engine model.^[8] Concurrently, Michio Sugeno introduced fuzzy integrals in his 1974 doctoral thesis, providing tools for non-additive measures in decision-making under uncertainty.^[9] The 1980s marked growing applications in control systems, building on Mamdani's work, while the 1990s witnessed integrations with emerging computational paradigms, including neural networks to form fuzzy neural systems for adaptive learning and genetic algorithms for optimizing fuzzy rules in complex optimization problems.^[10] Major contributors included George Klir and Bo Yuan, whose 1995 book Fuzzy Sets and Fuzzy Logic synthesized theoretical foundations and advanced fuzzy measures for uncertainty representation.^[11] By the 2020s, fuzzy mathematics has been incorporated into artificial intelligence, particularly through fuzzy neural networks for handling uncertainty in deep learning models and type-2 fuzzy sets for managing higher-order vagueness in big data analytics, with refinements rather than paradigm shifts since 2000.^[12] In 2025, the field celebrated the 60th anniversary of Zadeh's seminal paper through dedicated conferences, such as the IEEE International Conference on Fuzzy Systems.^[13] Early challenges in the 1970s, including criticism that fuzzy approaches lacked mathematical rigor and were dismissed as "not mathematics" by some academics, were overcome through empirical successes in engineering, leading to institutional recognition via the IEEE Computational Intelligence Society's Fuzzy Systems Technical Committee in the early 1990s and the founding of the International Fuzzy Systems Association (IFSA) in 1984.^[14]^[15]

Core Concepts

Fuzzy Sets

A fuzzy set A on a universe X is formally defined as a mapping \mu_A: X \to [0,1], where \mu_A(x) denotes the grade of membership of element x \in X in A, ranging from 0 (no membership) to 1 (full membership). This formulation generalizes classical crisp sets, where membership is binary (0 or 1), to accommodate partial degrees of belongingness, enabling the representation of vague or imprecise concepts.^[16] Fuzzy sets exhibit key properties that extend those of crisp sets. A fuzzy set A is normal if \sup_{x \in X} \mu_A(x) = 1, meaning it attains full membership for at least one element. It is convex if, for all x, y \in X and \lambda \in [0,1], \mu_A(\lambda x + (1-\lambda)y) \geq \min(\mu_A(x), \mu_A(y)) when X \subseteq \mathbb{R}^n, or equivalently, if all its \alpha-cuts are convex sets. Compactness holds if the support \{x \in X \mid \mu_A(x) > 0\} is a compact set. The \alpha-cut (or level set) of A for \alpha \in (0,1] is the crisp set A_\alpha = \{x \in X \mid \mu_A(x) \geq \alpha\}, which provides a way to decompose fuzzy sets into nested crisp subsets and facilitates proofs of many theoretical results.^[16] Basic operations on fuzzy sets A and B on the same universe X mirror classical set theory but use continuous functions on [0,1]. The union is defined by \mu_{A \cup B}(x) = \max(\mu_A(x), \mu_B(x)), the intersection by \mu_{A \cap B}(x) = \min(\mu_A(x), \mu_B(x)), and the complement by \mu_{\overline{A}}(x) = 1 - \mu_A(x). These operations are associative, commutative, and distributive, and satisfy De Morgan's laws, such as \overline{A \cup B} = \overline{A} \cap \overline{B}.^[16] For more flexible modeling, generalizations employ triangular norms (t-norms) for intersections, like the product t-norm \mu_A(x) \cdot \mu_B(x) = \mu_A(x) \mu_B(x), and triangular conorms (t-conorms) for unions, such as the probabilistic sum \mu_A(x) + \mu_B(x) - \mu_A(x) \mu_B(x). These allow adaptation to specific applications while preserving monotonicity and boundary conditions.^[17] Zadeh's extension principle provides a method to propagate fuzzy sets through functions, bridging crisp mathematics and fuzzy domains. For a crisp function f: X \to Y and fuzzy set A on X, the image fuzzy set f(A) on Y has membership function

\mu_{f(A)}(y) = \sup \{ \mu_A(x) \mid x \in X, f(x) = y \}.

If no such x exists, \mu_{f(A)}(y) = 0. This principle ensures that fuzzy inputs yield fuzzy outputs consistently with classical function behavior on \alpha-cuts. For example, consider a fuzzy set representing "young age" on X = \mathbb{N}, with \mu(20) = 1, \mu(40) = 0.5, and \mu(60) = 0; applying a function like "double the age" via the extension principle yields a fuzzy set for "young doubled age" with memberships adjusted accordingly. Under conditions such as f being continuous and \alpha-cuts of A being closed, the extension principle preserves continuity, ensuring the resulting fuzzy set is upper semicontinuous.^[18]

Membership Functions

In fuzzy set theory, the membership function \mu_A(x) plays a central role by quantifying the degree to which an element x belongs to a fuzzy set A, thereby capturing the inherent vagueness or imprecision in natural language concepts or real-world phenomena. Unlike crisp sets, where membership is binary (0 or 1), \mu_A(x) maps elements from the universe of discourse X to the interval [0, 1], with 0 indicating no membership, 1 full membership, and intermediate values partial membership. This function is typically chosen based on domain expertise, empirical data, or optimization techniques to reflect the intended fuzziness, ensuring that it aligns with the context of the application such as control systems or decision-making.^[16] Common shapes for membership functions include triangular, trapezoidal, Gaussian, and sigmoid, each offering distinct advantages in modeling different types of vagueness. The triangular membership function is defined as:

\mu(x) = \max\left(\min\left(\frac{x - a}{b - a}, \frac{c - x}{c - b}\right), 0\right)

for parameters a < b < c, providing a simple linear representation suitable for uniform transitions. The trapezoidal function extends this with a flat peak:

\mu(x) = \max\left(\min\left(\frac{x - a}{b - a}, 1, \frac{d - x}{d - c}\right), 0\right)

for a < b \leq c < d, ideal for plateaus in membership degrees like linguistic terms with sustained high relevance. The Gaussian function, mimicking natural distributions, is given by:

\mu(x) = e^{-\frac{(x - c)^2}{2\sigma^2}}

where c is the center and \sigma controls the spread, offering smooth, bell-shaped curves for continuous data. The sigmoid function, useful for monotonic transitions, is:

\mu(x) = \frac{1}{1 + e^{-k(x - c)}}

with k determining steepness and c the crossover point. These shapes are selected for their computational efficiency and interpretability in fuzzy systems.^[19] Key parameters of membership functions include the support, the set where \mu(x) > 0; the core, where \mu(x) = 1; and the crossover point, where \mu(x) = 0.5, which collectively define the function's extent and transition characteristics. In the context of linguistic variables—such as "small," "medium," or "large"—these parameters enable the representation of qualitative terms as fuzzy subsets over a numerical universe, facilitating approximate reasoning in fuzzy logic. Membership functions are often associated with such linguistic labels to model human-like decision processes.^[16]^[20] Selection of membership functions can employ vertical methods, which assign degrees pointwise via \mu(x), or horizontal methods, which define level sets through the inverse x(\mu), allowing for alternative representations that may simplify certain computations like interval arithmetic. In hybrid systems, parameters are optimized using techniques such as backpropagation in adaptive neuro-fuzzy inference systems (ANFIS), where gradients adjust shapes to minimize error against training data. Desirable properties include monotonicity (non-decreasing or non-increasing behavior) and unimodality (a single peak for convexity), ensuring intuitive and consistent fuzziness modeling; however, challenges arise with resolution, as too few functions per variable lead to coarse granularity, while excessive numbers increase computational complexity without proportional benefits.^[21] For instance, consider a fuzzy set for the linguistic term "hot" in temperature control, modeled with a trapezoidal membership function: \mu(t) = 0 for t < 30^\circC, linearly increasing to 1 at $40^\circC, remaining flat until $50^\circC, then linearly decreasing to 0 at $60^\circC. This design captures a range of comfortably high temperatures with full membership in the core [40, 50] and partial beyond, based on expert-defined thresholds for thermal comfort applications.

Fuzzification and Defuzzification

Fuzzification is the process of transforming a crisp input value into a fuzzy set by determining the degree of membership of that value to various linguistic terms using predefined membership functions. This step allows fuzzy systems to handle imprecise or uncertain inputs by mapping them to degrees between 0 and 1 across relevant fuzzy sets. Common methods for fuzzification include the singleton approach, where the membership function assigns a degree of 1 exactly at the crisp input value x and 0 elsewhere, which is computationally efficient for precise measurements.^[22] Gaussian fuzzification employs a bell-shaped membership function defined as \mu(x) = e^{-\frac{(x - c)^2}{2\sigma^2}}, where c is the center and \sigma is the standard deviation, providing smooth transitions suitable for continuous data.^[23] Piecewise linear methods, such as triangular or trapezoidal functions, offer simple, interpretable mappings that approximate gradual changes in linguistic variables. In rule-based fuzzy systems, the effective membership degree for an input across linguistic terms is often computed as the maximum over applicable rules or a weighted sum, enabling the aggregation of partial matches before inference.^[24] Defuzzification reverses this process by converting the aggregated fuzzy output set into a single crisp value for practical use, typically after fuzzy inference in control systems. Key methods include the centroid (center of gravity) approach, which calculates the crisp output y as y = \frac{\int y \mu(y) \, dy}{\int \mu(y) \, dy}, representing a balance point of the fuzzy distribution.^[25] The bisector method selects the value that divides the area under the membership function into two equal parts. The mean of maxima (MOM) computes the average of all points where the membership reaches its maximum, while the largest of maximum (LM) chooses the highest such point.^[26] For example, consider a speed control system where a crisp input of 60 km/h is fuzzified to a membership of 0.8 in the "fast" set and 0.2 in the "medium" set using triangular membership functions; after inference, the resulting output fuzzy set for throttle adjustment is defuzzified via the centroid method to yield a precise control signal, such as 75% throttle.^[27] Defuzzification methods are evaluated based on criteria like output accuracy in replicating desired behaviors and computational cost, with the centroid method being the most widely adopted for its reasonable balance in real-time applications.^[28] Challenges arise with non-convex output sets, which can exhibit multiple local modes, potentially leading to ambiguous crisp values that require additional heuristics for resolution.^[29]

Mathematical Structures

Fuzzy Logic

Fuzzy logic extends classical two-valued logic to accommodate partial truths and degrees of uncertainty by assigning truth values to propositions in the interval [0,1], where 0 represents complete falsity and 1 complete truth. This framework, rooted in fuzzy set theory, allows for the modeling of vague or imprecise statements, such as "the temperature is warm," by associating them with membership degrees rather than binary true/false assignments. Developed primarily by Lotfi A. Zadeh in the 1970s, fuzzy logic provides a mathematical structure for approximate reasoning, enabling the handling of linguistic variables and rules in systems where exactness is impractical.^[30] At its core, fuzzy logic operates on fuzzy propositions whose truth values are fuzzy sets over [0,1]. Logical connectives are defined using triangular norms (t-norms) and t-conorms to generalize classical operations: the conjunction (AND) can be the minimum (min(a,b)) or algebraic product (a·b), disjunction (OR) the maximum (max(a,b)) or probabilistic sum (a + b - a·b), and negation (NOT) as 1 - μ for a truth value μ. Implication operators, crucial for rule-based inference, include the Mamdani implication, defined as min(antecedent, consequent), and the Larsen implication, using the product antecedent · consequent. These connectives ensure monotonicity and boundary conditions, aligning with the properties of t-norms. Fuzzy logic draws from multi-valued logics, extending systems like Gödel's three-valued logic and Łukasiewicz's infinite-valued logic, which already allowed intermediate truth degrees beyond binary propositions. Inference in fuzzy logic employs the compositional rule of inference, where the output fuzzy set B' is obtained as the composition A' ∘ R, with A' the input fuzzy set and R the fuzzy relation representing the rule; for max-min composition, this is sup_x min(μ_{A'}(x), sup_y min(μ_R(x,y), μ_B(y))). This rule generalizes classical inference to fuzzy antecedents and consequents.^[30] A key inference mechanism is fuzzy modus ponens, which approximates the classical "if A then B; A is true, therefore B" for fuzzy inputs: given rule "if A then B" (relation R) and input A' ≈ A, the output is B' ≈ A' ∘ R. For instance, consider the rule "if temperature is hot then fan speed is high," where hot and high are fuzzy sets; if the observed temperature has membership 0.7 in hot, the inferred fan speed membership in high becomes 0.7 under min implication and max-min composition. This enables gradual activation of consequents based on input degrees. Certain theorems underpin the consistency of fuzzy logic operations under t-norms, including modularity (for implications I and t-norm T, I(x, T(y,z)) = T(I(x,y), z) when x ≥ y) and distributivity (T(I(x,y), z) = I(T(x,z), T(y,z)) for some families like Łukasiewicz t-norm). These properties ensure that fuzzy connectives behave associatively and distributively in inference chains, supporting reliable multi-rule systems.^[31]

Fuzzy Numbers and Arithmetic

Fuzzy numbers represent an extension of crisp real numbers to incorporate uncertainty and imprecision, defined as normal and convex fuzzy sets on the real line \mathbb{R} with bounded support and a membership function that is at least upper semicontinuous.^[32] This structure ensures that the fuzzy number captures a range of possible values with varying degrees of membership, where normality implies the existence of at least one element with full membership (height 1), convexity guarantees that the set is "bell-shaped" without dents, and bounded support limits the domain to a finite interval.^[33] The concept was formalized in seminal work by Dubois and Prade, who established fuzzy numbers as a foundational tool for handling vague quantities in mathematical modeling.^[34] A common and computationally tractable example of a fuzzy number is the triangular fuzzy number, denoted as (a, b, c) where a \leq b \leq c, with membership function given by:

\mu_{(a,b,c)}(x) = \begin{cases} 0 & x < a \text{ or } x > c \\ \frac{x - a}{b - a} & a \leq x \leq b \\ \frac{c - x}{c - b} & b \leq x \leq c \end{cases}

or equivalently, \mu_{(a,b,c)}(x) = \max\left(\min\left(\frac{x - a}{b - a}, \frac{c - x}{c - b}\right), 0\right).^[35] This form is widely adopted due to its simplicity and interpretability, often representing an approximate value b with symmetric or asymmetric spreads (b - a) and (c - b).^[36] Arithmetic operations on fuzzy numbers are primarily defined using Zadeh's extension principle, which extends crisp operations to fuzzy sets by preserving membership degrees. For two fuzzy numbers A and B, the membership function of their sum A + B is \mu_{A+B}(z) = \sup_{x + y = z} \min(\mu_A(x), \mu_B(y)), with analogous definitions for subtraction, multiplication, and division using the respective crisp operations in the supremum.^[33] Dubois and Prade demonstrated that these operations yield valid fuzzy numbers under the standard conditions, enabling the modeling of imprecise calculations such as in engineering tolerances.^[34] However, exact computation via the extension principle can be intensive, often approximated using \alpha-cuts, where the \alpha-cut of a fuzzy number A, denoted [A]_\alpha = \{x \in \mathbb{R} \mid \mu_A(x) \geq \alpha\}, is a closed interval for \alpha > 0. Operations on \alpha-cuts follow interval arithmetic, e.g., [A + B]_\alpha = [A]_\alpha + [B]_\alpha = \{x + y \mid x \in [A]_\alpha, y \in [B]_\alpha\}, which reconstructs the resulting fuzzy number by varying \alpha \in [0,1].^[37] This method improves efficiency while preserving convexity and normality.^[38] Fuzzy arithmetic exhibits certain properties that differ from classical real arithmetic; for instance, addition and multiplication are associative and commutative, but subtraction is neither, as (A - B) - C \neq A - (B + C) in general due to the non-invertibility of fuzzy subtraction under the extension principle.^[39] To compare or order fuzzy numbers, ranking methods are employed, such as the centroid method, which defuzzifies a fuzzy number A to its centroid \bar{x}_A = \frac{\int x \mu_A(x) \, dx}{\int \mu_A(x) \, dx} and compares these crisp values.^[40] For example, consider the triangular fuzzy number representing "3 ± 1" as (2, 3, 4) and "5 ± 0.5" as (4.5, 5, 5.5); their sum via \alpha-cuts approximates the triangular fuzzy number (6.5, 8, 9.5), illustrating how uncertainty propagates additively.^[41] Specialized variants include intuitionistic fuzzy numbers, which extend standard fuzzy numbers by incorporating both membership \mu and non-membership \nu functions satisfying \mu(x) + \nu(x) \leq 1, allowing for hesitation in addition to fuzziness.^[42] These were developed as part of intuitionistic fuzzy set theory to model scenarios with explicit refusal degrees, maintaining similar convexity and normality requirements on \mathbb{R}.^[43] Despite their utility, exact fuzzy arithmetic via the extension principle incurs computational complexity of O(n^2) when discretized over n membership levels, due to the pairwise evaluations in the supremum, motivating approximations like \alpha-cut methods for practical applications.^[44]

Fuzzy Relations and Topology

Fuzzy relations generalize classical binary relations by allowing degrees of association between elements, represented by a membership function \mu_R: X \times Y \to [0,1], where \mu_R(x,y) indicates the strength of the relation R between x \in X and y \in Y.^[45] This framework extends crisp relations, where membership is binary (0 or 1), to capture partial or graded connections. The composition of two fuzzy relations R \subseteq X \times Y and S \subseteq Y \times Z is defined using the max-min rule: \mu_{R \circ S}(x,z) = \sup_{y \in Y} \min(\mu_R(x,y), \mu_S(y,z)), which aggregates the strongest possible chain of associations through intermediate elements.^[45] Key properties of fuzzy relations mirror their classical counterparts but incorporate gradations. Reflexivity requires \mu_R(x,x) = 1 for all x \in X, indicating full self-association. Symmetry holds if \mu_R(x,y) = \mu_R(y,x) for all x,y, while transitivity is satisfied when \sup_{y \in X} \min(\mu_R(x,y), \mu_R(y,z)) \leq \mu_R(x,z) for all x,z. A fuzzy equivalence relation combines reflexivity, symmetry, and transitivity, partitioning sets into fuzzy classes of similarity. Similarly, a fuzzy partial order is reflexive, antisymmetric (if \mu_R(x,y) > 0 and \mu_R(y,x) > 0 then x = y), and transitive, enabling graded hierarchies.^[45] Fuzzy topology extends topological structures to fuzzy sets, defining a fuzzy topological space (X, \tau) where \tau is a collection of fuzzy open sets on X closed under arbitrary unions and finite intersections, with constant fuzzy sets 0 and 1 in \tau. This satisfies Kuratowski's closure axioms in a fuzzy context: the empty set (membership 0) and full set (membership 1) are open, unions of open fuzzy sets are open via pointwise supremum, and finite intersections are open via pointwise infimum. A basis for \tau consists of fuzzy open sets such that every open set is a union of basis elements, while a subbasis generates the topology through finite intersections forming the basis. Such spaces exist and generalize classical topology, preserving properties like continuity of fuzzy functions defined by preimage preservation of open sets.^[46] Fuzzy metrics introduce distance with imprecision, defined on a space (X, M, *) where M: X \times X \times (0,\infty) \to [0,1] satisfies axioms analogous to metric spaces: M(x,y,t) = 1 for t > 0 if x = y, strict triangle inequality M(x,y,s+t) \geq \min(M(x,z,s), M(z,y,t)) using a continuous t-norm *, and left-continuity in t. Here, M(x,y,t) represents the truth degree that the distance between x and y is less than or equal to some value at "time" or reliability t. This induces a topology on X via open balls \{y \in X \mid M(x,y,t) > \alpha\} for \alpha \in (0,1), \ t > 0.^[47] In examples, a fuzzy partial order models preferences where \mu_R(x,y) quantifies how much x is preferred over y, aiding decision-making under uncertainty. Fuzzy connectedness, defined via transitive closure of affinity relations in image spaces, measures pixel cohesion for segmentation, where higher degrees indicate stronger object belonging.^[45] Theorems confirm the robustness of these structures: fuzzy topological spaces satisfying Kuratowski axioms exist and support compactness notions, such as every fuzzy open cover having a finite subcover with positive infimum membership. Separation axioms adapt classically; a fuzzy T0 space distinguishes points by distinct neighborhoods with membership differences, while T1 requires singleton fuzzy closures to be closed (membership ≤ all points). These hold in induced fuzzy topologies from metrics.^[46] In pattern recognition, fuzzy relations underpin clustering algorithms that group data via transitive similarity, enhancing robustness to noise.

Applications and Extensions

In Control and Decision Systems

Fuzzy control systems apply fuzzy logic to manage complex, nonlinear processes where precise mathematical models are unavailable or impractical. These systems typically consist of a rule base derived from expert knowledge, such as "if error is high and change in error is positive then output is medium," followed by fuzzy inference to aggregate rule outputs and defuzzification to produce a crisp control signal.^[48] The Mamdani fuzzy controller, introduced in 1975, uses fuzzy sets for both antecedents and consequents, making it intuitive for linguistic rule representation and widely adopted for its interpretability in real-time applications. In contrast, the Sugeno or Takagi-Sugeno (TS) model, proposed in 1985, employs crisp functions in the consequents, often linear, which facilitates analytical design and integration with classical control methods. TS fuzzy systems blend fuzzy rules with local linear models, yielding an output computed as y = \frac{\sum_{i=1}^r w_i f_i(\mathbf{x})}{\sum_{i=1}^r w_i}, where w_i is the firing strength of the i-th rule, f_i(\mathbf{x} is a linear function, and r is the number of rules; this structure enhances computational efficiency and stability analysis. Early demonstrations include the 1975 Mamdani application to a steam boiler, marking the inception of fuzzy control, while the inverted pendulum—a benchmark for balancing unstable dynamics—has been stabilized using Mamdani rules since 1987, as demonstrated by Takeshi Yamakawa, showcasing fuzzy methods' ability to handle underactuated systems without full state feedback.^[49] Fuzzy PID controllers extend classical proportional-integral-derivative schemes by dynamically adjusting gains via fuzzy rules, offering superior performance in nonlinear environments, such as reduced overshoot and faster settling times compared to fixed-gain PID, particularly under parameter variations or disturbances.^[50]^[51] In decision systems, fuzzy mathematics enables multi-criteria decision making (MCDM) by incorporating uncertainty into weighting and ranking processes. Fuzzy analytic hierarchy process (AHP) assigns fuzzy weights to criteria hierarchies, allowing pairwise comparisons to reflect imprecise judgments, as in evaluating combined cooling-heating-power systems where linguistic scales quantify trade-offs between cost, efficiency, and emissions. Fuzzy data envelopment analysis (DEA) extends efficiency measurement to fuzzy inputs/outputs, transforming crisp DEA models into interval-based assessments to rank decision-making units under vagueness, with seminal work in 1998 proposing possibility and necessity measures for fuzzy observations. These approaches provide robustness in optimization tasks, such as resource allocation, by avoiding binary efficiency scores. Advantages of fuzzy control and decision systems include handling nonlinearities without exact models—relying instead on heuristic rules—and inherent robustness to noise and model mismatches, as evidenced by improved tracking in variable-road-condition simulations.^[50] Case studies highlight practical impact: fuzzy logic in automotive anti-lock braking systems (ABS) modulates brake pressure based on slip ratio and vehicle speed, enhancing stopping distance by 10-20% on low-friction surfaces while maintaining stability, as demonstrated in 2001 simulations for improved directional control.^[52] In medical diagnosis, fuzzy expert systems process symptom degrees of membership to infer conditions like multiple sclerosis, achieving over 90% accuracy in empirical tests by emulating physician reasoning under uncertainty.^[53] System performance is evaluated through metrics like Lyapunov stability, where fuzzy Lyapunov functions ensure global asymptotic stability by verifying negative definiteness of time derivatives along system trajectories, particularly for TS models. Tuning often employs genetic algorithms to optimize membership functions and rules, minimizing error criteria such as integral time absolute error, as in 1995 methods that evolved fuzzy controllers to mimic reference behaviors with convergence in under 100 generations for benchmark plants.

In Artificial Intelligence and Data Analysis

Fuzzy clustering plays a pivotal role in artificial intelligence by enabling the handling of data points that may belong to multiple clusters with varying degrees of membership, which is particularly useful in machine learning tasks involving overlapping or ambiguous data distributions. The Fuzzy C-Means (FCM) algorithm, a cornerstone of this approach, iteratively optimizes cluster centers and membership degrees to minimize an objective function that incorporates fuzziness. Specifically, FCM minimizes

J = \sum_{i=1}^c \sum_{k=1}^n \mu_{ik}^m \| x_k - c_i \|^2

where c is the number of clusters, n is the number of data points, \mu_{ik} is the membership of point x_k in cluster i, c_i is the center of cluster i, and m > 1 (typically m=2) controls the degree of fuzziness.^[54] This formulation, originally proposed by Dunn and refined by Bezdek, allows for soft assignments that better model real-world uncertainties compared to hard clustering methods. In AI applications, FCM facilitates unsupervised learning in scenarios like pattern recognition and anomaly detection, where data boundaries are not sharply defined. Neuro-fuzzy systems further integrate fuzzy mathematics with neural networks to enhance AI's ability to approximate complex functions under uncertainty. The Adaptive Neuro-Fuzzy Inference System (ANFIS), a hybrid model, combines the interpretability of fuzzy if-then rules with the learning capability of backpropagation in multilayer feedforward networks. ANFIS employs a Takagi-Sugeno fuzzy inference structure where premise parameters (membership functions) and consequent parameters (linear combinations) are tuned simultaneously using hybrid learning algorithms, such as least squares and gradient descent. This approach excels in tasks like time-series prediction and classification, providing both accuracy and linguistic explainability for AI models. In data analysis, fuzzy mathematics addresses imprecision in databases through techniques like fuzzy data mining, which leverages possibility theory to manage vague or incomplete information. Possibility theory, an extension of fuzzy sets, represents uncertainty via possibility distributions that distinguish between plausible and implausible values, enabling robust querying and pattern discovery in imprecise datasets. For instance, missing values can be handled by assigning possibility degrees rather than probabilistic estimates, improving the reliability of association rule mining in uncertain environments.^[55] This is particularly valuable in big data analytics, where traditional methods falter due to noise or vagueness. Practical examples illustrate fuzzy mathematics' impact in AI. In image segmentation, FCM partitions pixels into regions by assigning fuzzy memberships based on intensity similarities, effectively delineating overlapping tissues in medical imaging like MRI scans, as demonstrated in applications for brain tumor detection.^[56] Similarly, recommender systems employ fuzzy similarity measures to capture nuanced user preferences, such as computing fuzzy distances between item features to generate personalized suggestions that account for subjective ratings.^[57] To model deeper uncertainties, such as variability in membership functions themselves, type-2 fuzzy sets extend standard fuzzy sets by incorporating a footprint of uncertainty, allowing AI systems to handle dynamic or linguistic ambiguities more effectively. In interval type-2 fuzzy sets, the secondary membership grade represents the uncertainty in the primary grade, enabling better performance in noisy AI environments like natural language processing. Integration with deep learning enhances fuzzy mathematics' role in modern AI. Fuzzy activation functions, such as parameterized ones that emulate fuzzy logic operations (e.g., AND/OR), replace traditional activations like ReLU in neural networks to learn interpretable fuzzy expressions during training.^[58] For explainable AI, fuzzy rules provide human-readable justifications for model decisions; for example, Mamdani-type fuzzy inference systems generate linguistic outputs that trace predictions back to input features, improving trust in black-box models.^[59] Regarding performance, FCM outperforms k-means in datasets with overlapping clusters by allowing partial memberships, achieving higher accuracy in validation metrics like the Davies-Bouldin index for such cases, though its computational complexity is O(n c² d) per iteration due to membership updates and distance calculations—scalable variants like kernel FCM mitigate this for large-scale AI applications.^[60]

Advanced Topics and Variants

Type-2 fuzzy sets extend the foundational type-1 fuzzy sets by incorporating a secondary membership function that models uncertainty in the primary membership grades, allowing for a more nuanced representation of vagueness in membership shapes.^[61] In interval type-2 fuzzy sets, a common variant, the secondary memberships form an interval, which simplifies computations while capturing footprint-of-uncertainty effects, particularly useful for handling linguistic uncertainties in rule-based systems. This structure enables better minimization of uncertainties compared to type-1 sets, as demonstrated in applications requiring adaptive inference. Intuitionistic fuzzy sets, introduced by Atanassov, generalize fuzzy sets by including both a membership degree μ and a non-membership degree ν for each element, satisfying the condition μ + ν ≤ 1, with the remainder representing hesitation or indeterminacy.^[42] This formulation allows for a more comprehensive modeling of incomplete information, where non-membership is not simply the complement of membership, enhancing decision-making under conflicting evidence.^[42] The framework has been widely adopted for multi-criteria analysis due to its ability to balance affirmative and negative aspects explicitly.^[62] Fuzzy measures provide a non-additive generalization of classical measures, characterized by monotonicity rather than additivity, enabling the modeling of interactions among events in uncertain environments.^[63] The Sugeno integral, a key operation on fuzzy measures, is defined as ∫ f dμ = sup_α (α ∧ μ{f ≥ α}), where f is a measurable function, μ is the fuzzy measure, α ranges over [0,1], and ∧ denotes the minimum.^[64] This integral offers a lattice-theoretic alternative to the Choquet integral, prioritizing qualitative aggregation over linear combinations, and is particularly effective for ordinal data fusion.^[63] Rough-fuzzy hybrids integrate rough set theory's approximation capabilities with fuzzy sets to form granular computing frameworks, where rough approximations handle indiscernibility while fuzzy memberships address gradation.^[65] By combining Dempster-Shafer evidence theory with fuzzy granulation, these hybrids enable multilevel uncertainty representation through information granules, improving robustness in knowledge discovery from noisy data.^[66] This approach facilitates scalable processing in domains requiring both symbolic and subsymbolic reasoning.^[67] Quantum fuzzy logic merges fuzzy reasoning with quantum principles, incorporating superposition to represent entangled uncertainties in quantum computing paradigms.^[68] This variant allows fuzzy operations on qubits, enabling parallel evaluation of membership degrees across superposed states, which enhances computational efficiency for optimization problems under quantum constraints.^[69] Seminal work has focused on synthesizing quantum circuits for fuzzy relational inferences, bridging classical fuzzy logic with quantum gates.^[69] Current research in fuzzy mathematics emphasizes fuzzy big data analytics, where scalable fuzzy clustering and aggregation handle voluminous, heterogeneous datasets to extract patterns amid noise and incompleteness.^[70] Additionally, explainable fuzzy models have gained traction post-2018 under GDPR requirements, incorporating interpretable rule bases and visualization techniques to ensure transparency in automated decisions involving personal data.^[71] As of 2025, fuzzy mathematics continues to advance in hybrid systems, including fuzzy enhancements to large language models for uncertainty handling in natural language processing, as highlighted in recent conferences like Fuzz-IEEE 2025.^[13] Despite these advances, fuzzy sets face limitations in scalability for high-dimensional spaces, where the curse of dimensionality exacerbates computational complexity in rule generation and optimization.^[72] Ongoing work on type-n fuzzy sets addresses this by introducing higher-order memberships to hierarchically model escalating uncertainties, though practical implementations remain challenged by inference overhead.^[73] Looking ahead, integration of fuzzy mathematics with blockchain technology is emerging for managing uncertain transactions, using fuzzy consensus mechanisms to weigh probabilistic validations in decentralized ledgers.^[74] This hybrid approach promises enhanced security for ambiguous or partial data inputs in distributed systems.^[75]

References

[1]
[PDF] Fuzzy Systems Applications to Power Systems - UTK-EECS
Fuzzy mathematics is a broad field touching on nearly all traditional mathematical areas, the ideas presented in this discussion are intended to be ...
[2]
Fuzzy Mathematics - an overview | ScienceDirect Topics
Fuzzy sets are defined as collections of ordered pairs ((x, \mu_A(x))), where (x) is an element of a universal set (X) and (\mu_A(x)) is the membership function ...
[3]
None
### Summary of Fuzzy Mathematics from IEEE CAS 00.pdf
[4]
Fuzzy set theory - Zimmermann - Wiley Interdisciplinary Reviews
May 25, 2010 · In this review, the basic mathematical framework of fuzzy set theory will be described, as well as the most important applications of this theory to other ...Abstract · HISTORY · MATHEMATICAL THEORY... · APPLICATIONS
[5]
Fuzzy sets - ScienceDirect.com
A fuzzy set is a class of objects with a continuum of grades of membership. Such a set is characterized by a membership (characteristic) function.
[6]
[PDF] FUZZY-SET BASED LOGICS — AN HISTORY-ORIENTED ... - IRIT
The history of fuzzy logic starts with the foundational 1965 paper by Lotfi Zadeh entitled “Fuzzy Sets” [Zadeh, 1965]. In this paper, motivated by problems ...
[7]
Fuzzy Logic - Stanford Encyclopedia of Philosophy
Nov 15, 2016 · Fuzzy logic emerged in the context of the theory of fuzzy sets, introduced by Lotfi Zadeh (1965). ... Klir, 2017, Fuzzy Logic and Mathematics: A ...
[8]
An experiment in linguistic synthesis with a fuzzy logic controller
This paper describes an experiment on the “linguistic” synthesis of a controller for a model industrial plant (a steam engine)
[9]
Fuzzy Integrals | SpringerLink
Sugeno, M. [ 1974 ], Theory of Fuzzy Integrals and its Applications. Ph.D ... [1982], On fuzzy measure and fuzzy integral. In: Gupta and Sanchez ...
[10]
The merging of neural networks, fuzzy logic, and genetic algorithms
During the last decade, there has been increased use of neural networks (NNs), fuzzy logic (FL) and genetic algorithms (GAs) in insurance-related ...
[11]
[PDF] Fuzzy Sets and Fuzzy Logic: Theory and Applications
The first part of Fuzzy Sets and Fuzzy Logic provides a very carefully crafted introduction to the basic concepts and techniques of fuzzy set theory. The ...
[12]
A review on type-2 fuzzy neural networks for system identification - NIH
Mar 9, 2021 · Type-2 fuzzy neural networks (T2F-NNs) are extensively used in system identification problems, because of their strong estimation capability.Missing: 2020s | Show results with:2020s
[13]
Lotfi Zadeh and the Birth of Fuzzy Logic - IEEE Spectrum
Once the issue of classification had been solved, Zadeh could develop the theory of fuzzy sets quickly. Two weeks later he had a fairly fleshed-out group of ...
[14]
https://spectrum.ieee.org/lotfi-zadeh
[15]
Foundations of Fuzzy Logic - MATLAB & Simulink - MathWorks
Triangular membership function. trapmf — Trapezoidal membership function.
[16]
The concept of a linguistic variable and its application to ...
By a linguistic variable we mean a variable whose values are words or sentences in a natural or artificial language. For example, Age is a linguistic ...
[17]
The Design of Fuzzy Sets - IEEE Xplore
Vertical and horizontal schemes of membership estimation. Saaty's priority method of pairwise membership function estimation. Fuzzy sets as granular ...
[18]
The illustration of singleton vs. non-singleton FLS architecture
Non-singleton fuzzification is used to model uncertain (eg noisy) inputs within fuzzy logic systems.
[19]
gaussmf - Gaussian membership function - MATLAB
This function computes fuzzy membership values using a Gaussian membership function. You can also compute this membership function using a fismf object.
[20]
Mamdani Fuzzy Systems for Modelling and Simulation - JASSS
Mamdani fuzzy systems were originally designed to imitate the performance of human operators in charge of controlling certain industrial processes (Mamdani 1974 ...
[21]
Uncertainty Avoider Interval Type II Defuzzification Method - 2020
Jul 8, 2020 · The centroid defuzzification method used in fuzzy systems originated from mass centroid concept to find the centroid of a homogenous piece ...<|control11|><|separator|>
[22]
A survey of defuzzification strategies - Wiley Online Library
May 2, 2001 · Defuzzification is an important operation in the theory of fuzzy sets. It transforms a fuzzy set information into a numeric data information.Missing: seminal | Show results with:seminal
[23]
Application of fuzzy algorithms for control of simple dynamic plant
01 December 1974. Application of fuzzy algorithms for control of simple dynamic plant. Author: E.H. MamdaniAuthors Info & Affiliations. Publication ...
[24]
Defuzzification in Fuzzy Controllers - ACM Digital Library
We describe six important defuzzification methods and their respective merits and shortcomings, dependent on the rules, domains, etc. Further, we give an ...Missing: seminal | Show results with:seminal
[25]
Centroids of fuzzy sets when membership functions have spikes
Centroids are a common defuzzification method for fuzzy sets defined on the real line, and thus are practically important in type-1 and type-2 fuzzy logic ...Missing: origin | Show results with:origin
[26]
Many-Valued Logic - Stanford Encyclopedia of Philosophy
Apr 25, 2000 · Many-valued logic as a separate subject was created by the Polish logician and philosopher Łukasiewicz (1920), and developed first in Poland.
[27]
[PDF] Theory of T-norms and fuzzy inference methods
The triangular norm (T-norm) and the triangular conorm (T-conorm) origin- ated from the studies of probabilistic metric spaces [1, 2] in which triangular ...
[28]
Basic Concepts | SpringerLink
Basic definitions such as fuzzy subsets, fuzzy numbers, Zadeh's extension and operations of fuzzy numbers are presented. This chapter also introduces ...
[29]
[PDF] Operations on fuzzy numbers - Semantic Scholar
Jun 1, 1978 · The usual algebraic operations on real numbers are extended to fuzzy numbers by the use of a fuzzification principle, and the practical use ...
[30]
Operations on fuzzy numbers - Taylor & Francis Online
DIDIER DUBOIS ... In this paper, the usual algebraic operations on real numbers are extended to fuzzy numbers by the use of a fuzzification principle.
[31]
Triangular Fuzzy Number - an overview | ScienceDirect Topics
For illustrative purposes, we use a triangular fuzzy number (TFN), which is denoted A ̃ = ( l A , m A , u A ) , l A ≤ m A ≤ u A , where l A , m A , and u A are ...
[32]
[PDF] 5.3 Triangular Fuzzy Number
Dec 29, 2019 · Definition (Triangular fuzzy number) It is a fuzzy number represented with three points as follows : A = (a1, a2, a3) this representation is ...
[33]
[PDF] Arithmetic Operation of Fuzzy Numbers Using Α-Cut Method - IJISET
To calculate division of fuzzy numbers X and Y we first divide the α –cuts of X and Y using interval arithmetic. …………(4.4). To find the membership function ...
[34]
A computational method for fuzzy arithmetic operations
Oct 17, 2025 · In this paper, a computer implementation to evaluate the arithmetic operations on two fuzzy numbers with linear membership functions has ...
[35]
Some algebraic properties of weakly non-interactive fuzzy numbers
The authors treat the algebraic structure of fuzzy arithmetic based on the sup-(t-norm) convolution, and especially show some new properties comparing them ...
[36]
(PDF) On Fuzzy Arithmetic Operations: Some Properties and ...
We analyze a decomposition of the fuzzy numbers (or intervals) which seems to be of interest in the study of some properties of fuzzy arithmetic operations.
[37]
[PDF] ARITHMETIC OPERATION OF FUZZY NUMBERS BY USING 𝝈
In this article we have established some independent approach about the fuzzy arithmetic operations using 𝜎 − cut method. Keywords : Definition of fuzzy set, ...
[38]
Intuitionistic fuzzy sets - ScienceDirect.com
A definition of the concept 'intuitionistic fuzzy set' (IFS) is given, the latter being a generalization of the concept 'fuzzy set' and an example is described.
[39]
(PDF) Notes on triangular intuitionistic fuzzy numbers - ResearchGate
Aug 6, 2025 · Triangular IFN is a more generalized platform for expressing imprecise, incomplete, and inconsistent information when solving multi-criteria ...
[40]
[PDF] Fast Fuzzy A.-it, metic operations
Therefore, the computational complexity of computing the difference, product, or ratio of two fuzzy numbers is also O(nlog(n)).
[41]
An introductory survey of fuzzy control - ScienceDirect.com
This paper reviews studies on fuzzy control, key points in its application, recent industrial results, and important problems to be solved.
[42]
HUMAN ABILITY OF CONTROL AND ITS IMPROVEMENT BY ...
STABILIZATION OF DOUBLE INVERTED PENDULUM The inverted pendulum is essentially unstable and its control is not easy even when its dynamics can be ...
[43]
The Advantages of PID Fuzzy Controllers Over The Conventional ...
Fuzzy logic controllers (FLC's) have the following advantages over the conventional controllers: they are cheaper to develop, they cover a wider range of ...<|separator|>
[44]
Benefits of Intelligent Fuzzy Controllers in Comparison to Classical ...
Aug 30, 2023 · The PID controller, due to its robustness and simplicity, is widely employed in the industry; however, the algorithm has low performance when ...
[45]
[PDF] fuzzy logic controlled abs system for improved braking performance ...
Sep 7, 2001 · A fuzzy logic controlled Anti-Lock Braking System (ABS) is developed to improve the braking performance and directional stability during ...<|separator|>
[46]
Fuzzy C-means cluster analysis - Scholarpedia
Aug 26, 2009 · History. Fuzzy c-means clustering was first reported in the literature for a special case (m=2) by Joe Dunn ...Missing: paper | Show results with:paper
[47]
Fuzzy machine learning and data mining a
May 31, 2011 · In summary, fuzzy sets and related uncertainty formalisms, such as possibility theory, have made important contributions to the representation ...Machine Learning And Data... · Fuzzy Logic In Machine... · Potential Contributions Of...
[48]
Fuzzy C-Means Clustering: A Review of Applications in Breast ...
Jul 4, 2023 · This paper reviews the potential use of fuzzy c-means clustering (FCM) and explores modifications to the distance function and centroid initialization methods ...
[49]
Fuzzy Tools in Recommender Systems: A Survey
Mar 13, 2017 · This paper is ... Representation, similarity measures and aggregation methods using fuzzy sets for content-based recommender systems.
[50]
A parameterized activation function for learning fuzzy logic ... - arXiv
Aug 28, 2017 · A parameterized activation function for learning fuzzy logic operations in deep neural networks. Authors:Luke B. Godfrey, Michael S. Gashler.
[51]
Fuzzy inference system with interpretable fuzzy rules: Advancing ...
This review provides a fundamental understanding of interpretability and fuzzy rules, conducts comparative analyses of the use of fuzzy rules and other ...Missing: modern | Show results with:modern
[52]
[PDF] Fuzzy C-means and clustering algorithms: a comparative study
Fuzzy C-Means is more accurate than K-Means for overlapping clusters. In the second group of experiments, I will perform internal validation of Fuzzy C-.
[53]
Type-2 Fuzzy Logic: Theory and Applications - IEEE Xplore
Type-2 fuzzy sets are used for modeling uncertainty and imprecision in a better way. These type-2 fuzzy sets were originally presented by Zadeh in 1975.Missing: original | Show results with:original
[54]
Atanassov, K.T. (1986) Intuitionistic Fuzzy Sets. Fuzzy Sets and ...
Atanassov, K.T. (1986) Intuitionistic Fuzzy Sets. Fuzzy Sets and Systems, 20, 87-96. http://dx.doi.org/10.1016/S0165-0114(86)
[55]
[PDF] Fuzzy Measures and Fuzzy Integrals
This article gives a survey of the theory of fuzzy measures and fuzzy integrals. The measure is one of the most important concepts in mathematics and so.
[56]
The Sugeno integral. A point of view - ScienceDirect
The Japanese mathematician and engineer Michio Sugeno wrote the first paper concerning his integral in 1972 (see [23]). ... Beliakov et al. Discrete Fuzzy ...
[57]
Rough-Fuzzy Granular Computing, Case Based Reasoning and ...
Role of fuzzy-granulation (f-granulation) in machine and human intelligence, and its modeling through rough-fuzzy integration are discussed. Merits of fuzzy ...
[58]
Hybridization of Fuzzy Sets and Rough Sets - IEEE Xplore
Abstract: Fuzzy rough sets are the fruit of an intense and long-lasting collaboration effort between fuzzy set theory and rough set theory.
[59]
[PDF] Fuzzy-Rough Hybridization | Semantic Scholar
This paper highlights the potential of fuzzy quantifiers in creating new robust fuzzy rough models, advocates closer integration with granular computing as ...
[60]
Fuzzy Inference on Quantum Annealers - IEEE Xplore
Having the main components of a rule-based fuzzy logic system implemented on quantum computers, this paper provides an integrated solution for implementing a ...
[61]
On the quantum circuit implementation of fuzzy reasoning
Oct 2, 2025 · In this paper, quantum circuits for fuzzy relational inferences are synthesized. The compositional rule of inference (CRI) and ...
[62]
Editorial: Fuzzy Big Data-Driven Computational Intelligence Models ...
This special issue on “Fuzzy Big Data-Driven Computational Intelligence Models and Applications” aims to provide a comprehensive overview of the latest ...
[63]
(PDF) Paving the Way to Explainable Artificial Intelligence with ...
Aug 7, 2025 · Explainable Artificial Intelligence (XAI) is a relatively new approach to AI with special emphasis to the ability of machines to give sound ...
[64]
[PDF] Curse of Dimensionality for TSK Fuzzy Neural Networks - arXiv
Feb 8, 2021 · This paper explores why TSK fuzzy systems with Gaussian MFs may fail on high-dimensional inputs. After transforming defuzzification to an ...
[65]
(PDF) N-Fuzzy Sets and Their Applications - ResearchGate
The purpose of this paper is to define n-Fuzzy sets and study their relationship with intuitionistic fuzzy sets, Pythagorean fuzzy sets and Fermatean fuzzy sets ...Missing: ongoing | Show results with:ongoing
[66]
Fuzzychain: An equitable consensus mechanism for blockchain ...
Our proposal introduces an equitable consensus algorithm based on the proof of stake and fuzzy set theory to validate and verify block transactions, thus giving ...<|control11|><|separator|>
[67]
An Intuitionistic fuzzy cognitive mapping approach for blockchain ...
This study examines how blockchain technology (BT) enhances sustainability and resilience in construction supply chains (CSCs).